Maybe the question could be sharpened: "Can you tell about some examples where an IC in your company had a good feature idea, and what ended up happening?"
The problem there is that there's always ONE example. If you ask this in a Google interview, I guarantee they will tell you that "GMail started as a 20% project."
To me, a better signal is a set, repeatable process in place. Does it happen often enough that people know what to do?
When they tell you a story, they probably will minimize the gauntlet the innovator had to run.